A test of AI transparency developed by the Center for Research on Foundation Models, within HAI, gave every major company an F.